Nvidia 50 Series (Blackwell) support thread: How to get ComfyUI running on your new 50 series GPU.

[comfyanonymous]

I will try keeping this post up to date as much as possible with the latest developments.

To get your nvidia 50 series GPU working with ComfyUI you need a pytorch that has been built against cuda 12.8

In the next few months there will likely be a lot of performance improvements landing in pytorch for these GPUs so I recommend coming back to this page and updating frequently.

Windows

The recommended download is the latest standalone portable package or desktop installer that you can download from the README

Manual Install

If you install stable pytorch make sure it is cu128.

pytorch nightly cu128 is available for Windows and Linux:

pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

You can also use the Nvidia Pytorch Docker Container as an alternative which might give more performance.

Link: https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch

Here's how to use it:

docker run -p 8188:8188 --gpus all -it --rm nvcr.io/nvidia/pytorch:25.01-py3

Inside the docker container:

git clone https://github.com/comfyanonymous/ComfyUI
cd ComfyUI
grep -v 'torchaudio\|torchvision' requirements.txt > temp_requirements.txt
pip install -r temp_requirements.txt
python main.py --listen

[mcmonkey4eva]

There are also Windows prebuilt wheels of pytorch on cuda 128 that nvidia gave w-e-w to publish https://huggingface.co/w-e-w/torch-2.6.0-cu128.nv

[mmartial]

FYSA, the nightly for 12.8 appears available https://download.pytorch.org/whl/nightly/cu128

[comfyanonymous]

Yeah I see the package but it's still missing the cu128 torchvision, torchaudio and windows packages.

[mmartial]

It is. Thanks for clarifying the components we are waiting for.

I was trying to follow the guide by removing torchvision and torchaudio
and doing a pip3 install --pre torch --index-url https://download.pytorch.org/whl/nightly/cu128

spandrel too uses torchvision, which forces a downgrade on a Linux installation within a virtualenv.
https://github.com/chaiNNer-org/spandrel/blob/main/libs/spandrel/pyproject.toml#L36

EDIT: I will postpone testing until this is ready. Thanks for the details

[comfyanonymous]

I just noticed there are now nightly torchvision and torchaudio cu128 wheels for Linux. No windows packages yet though.

I tried them and they seem to work well so I have updated the post.

[mmartial]

Thank you 👍

With this news, I have made an updated release of https://github.com/mmartial/ComfyUI-Nvidia-Docker

For CUDA 12.8, I pre-install torch using pip3 install --pre torch torchaudio torchvision --index-url https://download.pytorch.org/whl/nightly/cu128 before installing ComfyUI from git and installing its requirements. I have tested it and can generate the bottle :)

No 50xx yet for me, so this test was done on a 3090 with driver 570 installed.

[edisonchan]

The Windows version said:
ModuleNotFoundError: No module named 'torchaudio'

[comfyanonymous]

This is because there was no torchaudio cu128 package provided.

The only thing this affects is the audio related nodes so it only affects you if you are using stable audio.

[edisonchan]

nf4 node does not work.

[ltdrdata]

Anything related to NF4 support should be resolved in the node pack repo that provides NF4 feature.

[edisonchan]

ImportError: tokenizers>=0.21,<0.22 is required for a normal functioning of this module, but found tokenizers==0.20.3.
Try: pip install transformers -U or pip install -e '.[dev]' if you're working with git main

How can I do to fix problem?

[adamreading]

Sorry - I have been banging my head against this for 6 hours - trying to run a cosmos workflow with 5090 - without SageAttention or TorchCompile - a 14 minute job on 4090 is taking 25 minutes. To turn them on I installed Triton and Sageattention - then I get this error afterwards whether I bypass the patch and compile or not:- KSampler
Command '['C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\bin\Hostx64\x64\cl.EXE', 'C:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k\cuda_utils.c', '/nologo', '/O2', '/LD', '/wd4819', '/IC:\Users\JoannaThompson\ComfyUI\ComfyUI_cu128_50XX\python_embeded\Lib\site-packages\triton\backends\nvidia\include', '/IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\include', '/IC:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k', '/IC:\Users\JoannaThompson\ComfyUI\ComfyUI_cu128_50XX\python_embeded\Include', '/IC:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\include', '/IC:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\shared', '/IC:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\ucrt', '/IC:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\um', '/FoC:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k\cuda_utils.cp312-win_amd64.obj', '/link', '/LIBPATH:C:\Users\JoannaThompson\ComfyUI\ComfyUI_cu128_50XX\python_embeded\Lib\site-packages\triton\backends\nvidia\lib', '/LIBPATH:C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\lib\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', '/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.42.34433\lib\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64', '/LIBPATH:C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64', 'cuda.lib', '/OUT:C:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k\cuda_utils.cp312-win_amd64.pyd', '/IMPLIB:C:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k\cuda_utils.cp312-win_amd64.lib', '/PDB:C:\Users\JOANNA1\AppData\Local\Temp\tmpy8t3476k\cuda_utils.cp312-win_amd64.pdb']' returned non-zero exit status 2.

I have run this through GPT and followed every instruction. Step 1: Verify CUDA Toolkit Installation
Run the following command in ComfyUI's embedded Python:

python
Copy
Edit
import torch
print(torch.version.cuda)
Expected Output: 12.8

If it does not match 12.8, you may need to reinstall PyTorch with the correct CUDA version.

✅ Step 2: Check NVIDIA Toolkit and Drivers
Ensure you have the latest NVIDIA GPU driver and CUDA toolkit installed.

Update your NVIDIA driver (Download latest)
Confirm CUDA installation with:
powershell
Copy
Edit
nvcc --version
Expected Output: release 12.8
If not, reinstall CUDA 12.8 from NVIDIA.

✅ Step 3: Fix Microsoft Visual Studio Build Tools
Open Developer Command Prompt for VS 2022

Run:

sh
Copy
Edit
cl.exe
If it says command not found, reinstall Microsoft C++ Build Tools:
Download: Visual Studio Build Tools 2022
Install:
C++ CMake tools for Windows
MSVC v143 - VS 2022 C++ x64/x86 build tools
Windows 10 SDK (latest version)
Restart the PC.

[yu-jingxiang]

your questiong in the same to me.

[rubentorresbonet]

anyone knows if other pytorch's cuda versions like 12.6 will work with blackwell?

[mpfcs162]

No, Blackwell GPUs require 12.7 or newer IIRC.
Edit: So 12.8 since that is the current version. It's a new architecture so I guess it makes sense that 12.6 and older versions are not supported. CUDA is not backwards compatible?

[fasfafefggwgrw]

When I use [ComfyUI package with a cuda 12.8 torch build], many custom_nodes are "IMPORT FAILED" including Manager, InstantID, ReActor ...

[Shad0w23333]

same, is is possible to build a PyTorch that works for 50s?

[mpfcs162]

Check Python version, reinstall dependencies, and look for incompatibilities in the logs - some nodes might need updates for CUDA 12.8. Fresh install or upgrade?

[dirtmonster1337]

i'm having the same issue, this package uses python 3.13 but ImpImporter was discontinued in these builds, anyone have a fix?

[Shad0w23333]

I will try keeping this post up to date as much as possible with the latest developments.

To get your nvidia 50 series GPU working with ComfyUI you need a pytorch that has been built against cuda 12.8

In the next few months there will likely be a lot of performance improvements landing in pytorch for these GPUs so I recommend coming back to this page and updating frequently.

Windows

Windows users can download this standalone ComfyUI package with a cuda 12.8 torch build

Manual Install

At this moment Blackwell is not yet supported by stable pytorch.

pytorch nightly cu128 is available for Linux:

pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

You can also use the Nvidia Pytorch Docker Container as an alternative which might give more performance.

Is is possible to build a pytorch that works for 5090? If so how to do it?

[adamreading]

What do you mean? There is already one for CUDA 12.8: https://download.pytorch.org/whl/nightly/cu128/torch/ For Windows you can try the EA version found here: https://huggingface.co/w-e-w/torch-2.6.0-cu128.nv Or just use the standalone package linked above with a CUDA 12.8 torch build.

We spent 6 hours + yesterday trying to install comfyui portable around the special version at the top of the screen. It wont let me run Sageattention or Torchcompile in workflows - and because it’s missing TorchAudio - things I use every day like GripTape won’t load into custom nodes any more. Running the cosmos workflow without sage and compile - is twice as long as on a 4090 system (25 mins for 8 sec gen compared to 14 minutes on 4090) Although I was able to install SageAttention and Triton I don’t think they are working with this version of PyTorch

Have you checked the logs for errors or tried a clean environment to rule out conflicts?

Yes we went through the logs - they are in the earlier post above - we tried for 6 hours downloading different versions - Visual build tools, various Nvida toolkits - did everything GPT suggested - its a brand new out of the box machine - nothing previously installed except office.

[adamreading]

5090 Crash 10.txt
One of the complete logs

[mpfcs162]

Tried running Linux? https://download.pytorch.org/whl/nightly/cu128/torch/
To my knowledge there is no official Windows build of PyTorch that supports 12.8 yet.
Sorry I can't offer more help...

[adamreading]

Tried running Linux? https://download.pytorch.org/whl/nightly/cu128/torch/ To my knowledge there is no official Windows build of PyTorch that supports 12.8 yet. Sorry I can't offer more help...

Thanks for looking. The app I want to run (a comfyui-discord bot) requires portable and lm
Studio to work apparently. Plus I use the machine for office365 at the same time lol. Looks like I just have to be patient and work out what will run and what won’t. It ran an ltxv video generator in 36 seconds so some stuff is ok. Couldn’t load anything Pulid.

[hellonearthis]

I got it working with the https://huggingface.co/w-e-w/torch-2.6.0-cu128.nv files.
I turned them into ZIP files which I opened. I then went to the folder ComfyUI.venv\Lib\site-packages and deleted the corresponding folders and extracted the zip folders there.

One of the corresponding files was like ...+cu126.dist-info and I deleted it and plopped in the ...+cu128.nv.dist-info file from the zip.

I'm sure there are better ways to do the but it worked for me.

[adamreading]

Having given up on Portable for now - too many errors lol - I am using WSL - Ubuntu - everything else is set up and working in Comfyui - I have the latest pytorch nightly from pip install --pre torch torchvision torchaudio --index-url[ https://download.pytorch.org/whl/nightly/cu128] - but when I use Sageattention I get this:
loaded completely 29493.675 3667.902587890625 True
0%| | 0/80 [00:00<?, ?it/s]'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
'sm_120' is not a recognized processor for this target (ignoring processor)
LLVM ERROR: Cannot select: intrinsic %llvm.nvvm.shfl.sync.bfly.i32
Aborted (core dumped)

[mpfcs162]

Wait, can u explain how to do that? Thank u

Do what?

[adamreading]

I managed to get it working in Windows yesterday. Had to manually install the torch-2.6.0+cu128.nv-cp312-cp312-win_amd64.whl and torchvision-0.20.0a0+cu128.nv-cp312-cp312-win_amd64.whl.

Originally had the wrong Python version installed. Easy fix but took me a long time (longer than I want to admit) to figure out...

There are some errors though (torchaudio missing), but I'm running SD 3.5 L to test on a 5090 and I get output. And fast too. But 32GB VRAM is not as much as I'd hoped and the impact compared to a 4090 is not that big - about ~30% faster which is expected I guess...

But I'd honestly recommend just running it on Linux atm.

Have you managed to get SageAttention working yet? (With Linux or Windows)

[mpfcs162]

No. I assume it won't be fully stable until we see an official release of PyTorch with CUDA 12.8 support. Or req some additional manual config. I just can't be bothered with that at this point. Been stuck in integration hell all day trying to troubleshoot some code for a client so haven't had much time to dig further into this topic. Sorry.

[merecesarchviz]

wrong phyton version!? witch version do you install to work? 3.12? I have the same errors that you and i have version 3.11

[Matriv-org]

Hi, Any update about it ?I had a working Sageattention2 and Torchcompile with 4090, but I've the same 'sm_120' is not a recognized processor for this target (ignoring processor) on 5090

SOLVED:
I've WAN2.1 working with torch compile + SageAttention2
In case you had installed Sageattention2 + Triton, you must update triton with
python_embeded\python.exe-m pip install -U --pre triton-windows
It works with torch 2.8.0.dev20250401+cu128, python3.12.9.

Here is the github: https://github.com/woct0rdho/triton-windows

[Teuflor]

im using comfyui over pinokio, is there any way to update my comfy to work with my new 5080?

i deleted the old files and changed in the torch.js on the patch above, but it wont start

[gachapengames]

Will using a docker allow for a working torchvision in windows?

[Six6pounder]

@gachapengames Yes, use nvcr.io/nvidia/pytorch:25.01-py3 as image and it works

[gachapengames]

Can someone write a little guide for getting a docker running torchvision, etc on window with the portable Blackwell comfyui release? Please write it for a normal user who has no programming knowledge.

There are basic instructions in the OP, but what even is a docker?

"docker run -p 8188:8188 --gpus all -it --rm nvcr.io/nvidia/pytorch:25.01-py3"

Where is this command supposed to be run? Is a docker something to be installed to system python or standalone folder?

How would this be installed on a fresh portable comfyui install?

With more Blackwell cards trickling out, there will most likely be more users needing help setting this up. Thank you.

[mmartial]

Please check the "Windows: WSL2 using podman" from https://github.com/mmartial/ComfyUI-Nvidia-Docker

With the caveat that you will need to adapt the podman run line if you want to use the proposed containers to be something like (adapted to attempt to reflect the above order):

podman run --userns=keep-id -p 8188:8188 --device nvidia.com/gpu=all --rm -it nvcr.io/nvidia/pytorch:25.01-py3

with the caveat that everything will be stored within the container since you have no mount (the -v from my original writeup)

so maybe:

podman run --userns=keep-id -p 8188:8188 --device nvidia.com/gpu=all --rm -it -v `pwd`/comfy:/comfy nvcr.io/nvidia/pytorch:25.01-py3

and then do the other steps listed above (pip install...) after cd /comfy so that the files persist when you exit the container

[gachapengames]

To anyone having trouble with the portable version for Blackwell gpus, do not update it! I noticed during updating that it uninstalled the cu128 and installed another version for example. Torchvision works with a fresh install without updating!

[mpfcs162]

So I've used both Comfy UI and Stability Matrix with SD Forge WebUI these past days since getting my 5090.

Each instance needs to have the correct version of Torch, and updating/installing will not ensure that (edit: if anyone reads this in the future there will eventually be an official torch release that supports CUDA 12.8 and in that case what is written here might no longer valid). Best way is to install Torch from the files above, making sure you use the correct version for your system and Python install. For SM, you need to install it to the package directly for example.

And please, do not run commands you find online unless you understand what they do. This applies to the below commands as well.

Navigate to the correct dir and type this to uninstall torch, torchvision and torchaudio:
python_embeded\python.exe -m pip uninstall -y torch torchvision torchaudio

Then, for ComfyUI at least, you can run this. Adjust the command as needed:
python_embeded\python.exe -m pip install H:\torch\torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl H:\torch\torchvision-0.20.0a0+cu128.nv-cp310-cp310-win_amd64.whl

You can run this in your dir to check what is installed:
python_embeded\python.exe -m pip list

[mpfcs162]

I just want to add that asking chatGPT (and similar LLMs) about this might not be as helpful as you think.

A very common issue is coherence with these models, and they can easily spew out answers that look and seem correct, while in reality are not what you actually need. For example, if I ask it why I'm getting kernel error when attempting to run stock ComfyUI it will tell me to update using a pip command, but it will give me a link to a build that does not exists (cu127) since it believes that CUDA 12.7 is the version I need for my 5090 since that is the min. version of CUDA that Blackwell supports.

I can follow that trail together with chatGPT for hours and end up trying to compile my own version, uninstall/reinstall various extensions and packages and get absolutely nowhere. And chatGPT will always act and respond with 100% confidence no matter what, and I will never figure out the culprit unless I go to the repo or follow the trail myself to discover that it (cu127) does not exists.

Of course, LLMs are great for a number of reasons - but beware when trying to do stuff like this that the output might be complete nonsense.

[adamreading]

I just want to add that asking chatGPT (and similar LLMs) about this might not be as helpful as you think.

A very common issue is coherence with these models, and they can easily spew out answers that look and seem correct, while in reality are not what you actually need. For example, if I ask it why I'm getting kernel error when attempting to run stock ComfyUI it will tell me to update using a pip command, but it will give me a link to a build that does not exists (cu127) since it believes that CUDA 12.7 is the version I need for my 5090 since that is the min. version of CUDA that Blackwell supports.

I can follow that trail together with chatGPT for hours and end up trying to compile my own version, uninstall/reinstall various extensions and packages and get absolutely nowhere. And chatGPT will always act and respond with 100% confidence no matter what, and I will never figure out the culprit unless I go to the repo or follow the trail myself to discover that it (cu127) does not exists.

Of course, LLMs are great for a number of reasons - but beware when trying to do stuff like this that the output might be complete nonsense.

I concur from personal experience - however as someone who doesn't understand this stuff - and having had an incredible machine arrive a week ago, that I just couldn't get to work - and all my techie friends on discord were at a loss as well - ChatGPT was my only option. Since talking to LLMS is what i do a lot for work etc, I just kept everything inside one thread - so it remembered everything we had been doing - I used its commands to trouble shoot - but used my brain to realise when we were repeating or when it had forgotten about having to use 12.8 sm_120 etc and kept reminding it of these things - and with a little prompting it helped me through a maze of tests and trials I never would have got through on my own. In the end - I went back to the original Comfy Windows stand alone install thats at the top of the page. I learned to install custom nodes in new workflows - 1 at a time, restarting and checking the logs after every node - and trouble shooting anything that threw up one step at a time. And it's now pretty much running. It is slow however - without SageAttention or FlashAttention etc I think it struggles. And I am sure there will be more performance increases as more tools get compiled to use its abilities. I am old and have been around a while - so I realise that life at the bleeding edge is not always easy. I am still confused about what the difference is between Comfyui Stand Alone and Comfyui Portable - half my friends are saying go one way the other the other - but I think they are just two different names for the same install? As far as I can tell (having installed all of these this week - for a Windows machine the options are 1) Desktop Comfy that installs from an exe file - doesn't open in a browser - has a built in terminal in the bottom pull up in a tab next to the logs, but seemed restrictive to use (though I can't tell you why). 2) The Windows Stand Alone I am currently using - which, once you figure out where to run the pip commands etc is pretty straight forward, easy to back up directories - or reinstall by dragging the zip file over the top again 3) WSL - allowed me to use TorchAudio right away - feels like gets quicker support from the community because its in Linux - but got quite complicated very quickly - and installing an unrelated program in a docker in the same WSL killed the whole thing - and I lost everything and no way to recover my files... @mpfcs162 a case in point for your statement - ChatGPT told me to remove the WSL to clear the issue with YuE and gave me the command to run - which I did - and then it wiped everything. (My own fault)

[edisonchan]

So I've used both Comfy UI and Stability Matrix with SD Forge WebUI these past days since getting my 5090.

Then, for ComfyUI at least, you can run this. Adjust the command as needed: python_embeded\python.exe -m pip install H:\torch\torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl H:\torch\torchvision-0.20.0a0+cu128.nv-cp310-cp310-win_amd64.whl

You can run this in your dir to check what is installed: python_embeded\python.exe -m pip list

python_embeded\python.exe --version
Python 3.12.8
PS E:\ComfyUI_cu128_50XX> python_embeded\python.exe -m pip install ".\update\torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl" ".\update\torchvision-0.20.0a0+cu128.nv-cp310-cp310-win_amd64.whl"
ERROR: torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl is not a supported wheel on this platform.

not work?

[adamreading]

So I've used both Comfy UI and Stability Matrix with SD Forge WebUI these past days since getting my 5090.
Then, for ComfyUI at least, you can run this. Adjust the command as needed: python_embeded\python.exe -m pip install H:\torch\torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl H:\torch\torchvision-0.20.0a0+cu128.nv-cp310-cp310-win_amd64.whl
You can run this in your dir to check what is installed: python_embeded\python.exe -m pip list

python_embeded\python.exe --version Python 3.12.8 PS E:\ComfyUI_cu128_50XX> python_embeded\python.exe -m pip install ".\update\torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl" ".\update\torchvision-0.20.0a0+cu128.nv-cp310-cp310-win_amd64.whl" ERROR: torch-2.6.0+cu128.nv-cp310-cp310-win_amd64.whl is not a supported wheel on this platform.

not work?

You are running python 3.12. You are trying to use the wheels for python 3.10… first go here and download the two 3.12 files to the same directory you are running the command in. https://huggingface.co/w-e-w/torch-2.6.0-cu128.nv Them adapt the command to use your version of the wheel you just downloaded.

[Tathiane123]

Excelente

[Tayyab-H]

Are there no windows pytorch options available for cuda 12.8 still? Any help would be greatly appreciated.

[adamreading]

Looks like you can import any model though - https://lmstudio.ai/docs/basics/import-model

[adamreading]

looks like 2gb of my gpu VRAM is there used all the time with nothing obviously running - then adding a 5gb Quen took it up to 7GB

[adamreading]

llava 1.5 used 12gb VRAM ran super fast

[Tayyab-H]

thanks for the help! Ill give LM studio a try! Its a shame it isnt working through code, id love to build stuff with it.

[pocketpoetry]

How did you get LMstudio to work on a 50 series card? mine will just refuse to load and give me this error, which I assume its cause it doesnt detect a CUDA "Failed to load LLM engine from path: C:\Users\user.lmstudio\extensions\backends\llama.cpp-win-x86_64-nvidia-cuda-avx2-1.22.0\llm_engine_cuda.node. Cannot find module 'C:\Users\user.lmstudio\extensions\backends\llama.cpp-win-x86_64-nvidia-cuda-avx2-1.22.0\llm_engine_cuda.node'
Require stack:

• C:\Users\user\AppData\Local\Programs\LM Studio\resources\app.webpack\lib\llmworker.js"

[logan683]

What is the current method of installing the manager without getting errors? I use ComfyUI portable Windows.

Thank you.

[pocketpoetry]

What is the current method of installing the manager without getting errors? I use ComfyUI portable Windows.

Thank you.

I follow this video to install mine without errors.. you can give ir a try

https://youtu.be/a-IXXa9P_RY?si=FW4-8MUgcRQFGper

[comfyanonymous]

The official standalone package has now been updated to pytorch 2.7 cu128 so just download that if you have a 50 series.

The desktop package will be updated in the next week or so assuming everything goes well.

[akedia]

pytorch 2.7 cu128 does not work the model compile node

[Jatts-Art]

After finally getting ComfyUI to work and installing what was needed... wtf? Generating basic images takes pretty much the same exact speed using RTX 5080 then it did for my RTX 4060? Is this normal?

[Jatts-Art]

You need to completely copy the entire log (Ctrl - A, Ctrl - V). The most interesting part comes next, and let the GPT chat analyze it.

After providing DeepSeek the entire output log and other information about what I have installed, they pointed out the following potential issues which I've simplified for listing here (as well as my own notes about my thoughts for each):

• Your driver: 576.02 is critically outdated for RTX 5080 (I installed these drivers literally this week)
• Recommended to perform DDU (umm... no?)
• Pytorch 2.8.0.dev + CUDA 12.8 is unstable and lack optimizations (lol wut)
• Flash-Attention not available + uninstall xformers (bruh)
• VRAM Allocation Mode: set --highvram --force-fp32 --deterministic (maybe this is it?)
• cudaMallocAsync can cause 20-30% slowdown on Ada/Lovelace/Blackwell GPUs (wat)
• Python v3.12.8 is not compatible, recommend v3.10.12
• ComfyUI Version is outdated (highly doubt this is the reason, personally)
• Impact-Pack Installation Corruption (this matters, even if Im on a workflow thats not using these nodes? Nah)
• Jovimetrix Node Failures
• Disable Power Saving features (they already are)

And there, thats about it lol... the only thing on here that makes any remote sense to me is the VRAM Allocation Mode? It is very important to not always rely on what AI says, in certain cases, cause some of this dont sound right to me at all quite honestly. Especially regarding the drivers and even saying I should perform DDU, lol...

[Turkinolith]

I feel strongly compelled to note that GPT is prone to hallucinations ant your comment about the driver definitely feels like one. Especially since this is all bleeding edge tech that is getting updated on the daily and GPT's info cutoff is likely not up to par.

[Jatts-Art]

I feel strongly compelled to note that GPT is prone to hallucinations ant your comment about the driver definitely feels like one. Especially since this is all bleeding edge tech that is getting updated on the daily and GPT's info cutoff is likely not up to par.

Yeah earlier in the week I was already using it to help me get started to begin with and one of its first responses was to say that the RTX 5080 "doesnt exist yet"

[Jatts-Art]

After yelling at and insulting the AI a bit more, it eventually helped refine a .bat file with the following details within (may as well share, to see what others think):

@echo off
set PYTHONPATH=
set CUDA_LAUNCH_BLOCKING=0
set PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.9,max_split_size_mb:1024

"C:\AI\ComfyUI-Portable\python_embeded\python.exe" -s "C:\AI\ComfyUI-Portable\ComfyUI\main.py" ^
--force-fp16 ^
--preview-method latent2rgb ^
--use-sage-attention ^
--normalvram ^
--fast fp16_accumulation ^
--disable-cuda-malloc ^
--windows-standalone-build

pause

It specifically mentioned that --highvram and --gpu-only are not good for Blackwell, but who knows if thats true... I'll see after testing various things today.

It also continues to highly recommend Flash-Attn but I cannot build/install that with my current version of Python/Torch nor do I know of where to even find the correct version to use with what I got... and also, according to other people I know, Flash-Attn barely impacts much compared to Sage-Attn anyway? No idea, but I'll test to see if any of this helps me.

[Jatts-Art]

Found something extremely vital for my problem, I think: #1065 (comment)

My startup CMD does NOT say "Using xformers attention" at all... and judging from what people keep saying, this is evidently very bad for speeds? Except that I do have it installed, based on the one shared here within this very same discussion page, thats meant to be compatible with the version of Python & Pytorch that I am using. I also do not get any errors/warnings regarding my xformers either, so whats the the deal?? I certainly hope its not because I am using Python v3.12.8 rather than v3.12.9? Bruh.

EDIT1: Okay it is due to --sage-attention which the AI kept gaslighting me into using, which was overwriting xformers. So I guess I cant use both? And if not, which is better to use... Sage-Attn or Xformers? Or should I try to somehow get Flash-Attn to installed?

EDIT2: I was doubly gaslit!! I think they actually can be used together, even if "Using xformers attention" isnt mentoned? This reddit thread mentions using both together, but does not mention anything at all about potentual conflicts: https://www.reddit.com/r/StableDiffusion/comments/1iztzbw/impact_of_xformers_and_sage_attention_on_flux_dev/

So, overall, now I am just completely confused on wtf is working and not working. AI chat dont help at all, other than making things more complicated/confusing; was horrible advice for people to recommend it here to begin with imho, and a bad habit for current generations to constantly rely on it. I think I'm done with that and will wait for a Human response here instead.

[psevdonimb]

Question for RTX 50 graphics card owners. Do you have the basic FLUX models working? Not the unet (nf4) models, not *.GGUF, but the ones called flux1-dev.safetensors or flux1-schnell.safetensors.

[Turkinolith]

Yes, I have been running the fp8 version of flux1-dev with no issues.

[psevdonimb]

Interesting... Do you really have the base model that is 23.8GB in size and downloaded from here?:
https://huggingface.co/black-forest-labs/FLUX.1-dev/tree/main

What operating system do you have?
My base models don't work anywhere, neither in ComfyUI nor in Forge and I started to think that the rtx 50xx video cards don't have support for this yet. I have the same thing with the base models of HiDream on Windows 10 with rtx 5090 (only *.gguf works)

[mcf100]

Yes, 5090, full flux models working (24GB one). Windows 11

[Rando855]

Followed a guide for the Windows portable for 50xx series cards and am receiving an error: Error running sage attention: PY_SSIZE_T_CLEAN macro must be defined for '#' formats, using pytorch attention instead.

Is this due to Triton 3.3.0? I read that it should be 3.2.0.

[Jatts-Art]

I am literally getting this same exact error and was just about to come on here and ask about it myself!

I get Error running sage attention: PY_SSIZE_T_CLEAN macro must be defined for '#' formats, using pytorch attention instead. spammed over and over, maybe thats whats causing my speeds to be shit as well, which ive been trying to figure out all day.

[Rando855]

It's definitely a new version issue. I'm on a clean install of Windows 11, clean install of Python 3.13, portable comfy, uninstalled and reinstalled everything numerous times. Same error

[Jatts-Art]

It's definitely a new version issue. I'm on a clean install of Windows 11, clean install of Python 3.13, portable comfy, uninstalled and reinstalled everything numerous times. Same error

Came across this, coincidently, after I mentioned the issue elsewhere.
https://www.reddit.com/r/StableDiffusion/comments/1jk2tcm/step_by_step_from_fresh_windows_11_install_how_to/mp1xvej/?context=3

[Baku-Rue]

Running into the same issue after attempting to do a clean install of the portable version. Was a clean install of Windows 11, and python 3.13 like the others. No clue how to go about trying to fix it though.

[gonzalu]

Hello everyone,

Thanks so far for all the great efforts by all on this thread. Thank you Comfy for your tireless work on this project.

I am not sure my question has been answered after going through pretty much the whole thread since the OP.

The install that I use every day is one that I installed from scratch using ComfyUI Portable about a year go (or more, I think January 2024)

I have updated and updated and updated almost once per day or multiple times per day via Manager ever since then. I just updated a few minutes ago. EVERYTHING works absolutely perfectly on my 4090.

Getting a 5090 is in my near future and I am terrified of breaking my current install or worse, starting from scratch yet again.

is there any solution for someone like me with an old Portable version that has been updated endlessly? I see that a fresh install "just works" but was hoping that there was a similar "run this batch file and it just works" from inside the updates directory :D

Thanks again everyone.

Cheers!

[pocketpoetry]

Honestly in my humble opinion.. coming from using a RTX 4050 laptop for 2 years until my recent upgrade to a 5070 ti PC

I totally understand the feeling of avoidance to start from scratch, infact I was scared of the change so much that I actually avoided installing a fresh Comfyui only RTX 5070 ti for 3 to 4 days after I got my new PC, instead i went back to use my 4050 laptop just to feel 'Safe', which is a totally illogical action.

But from my personal experience, the best way to learn and become better in this is to brave the front and face all the possible incompatibilities and errors that is certain to come to your face when u do a major change

The more you face it, the better you become in it, fear will become frustration, frustration will become irritance, and irritance will become just another minor occupancy, that's when u know nothing is a worry for you cause you know your stuff and is ready for any hardware upgrades without fear of incompatibilities

It's a tough but very fruitful journey to take, and only you can take the first step, chatgpt or any coding llms can only help you so far but they cannot be relied on.. but the rewards is astounding when u realise most installations are not that tough to understand.. but a hassle to go through

Good luck in your journey :)

[gonzalu]

Thanks for the wonderful reply.

[Mads-Q]

I installed the windows portable v.0.3.31 that was released yesterday: https://github.com/comfyanonymous/ComfyUI/releases/tag/v0.3.31
So far everything works flawless on my RTX5090! I used some of my favorite workflows for Flux and Wan2.1, after updating custom notes with comfy manager, all good. Yay.

This is from the release notes on github:
'Update portable package workflow to cu128 by @comfyanonymous in #7812'

[gonzalu]

Dumb question but, what benefit does Sage Attention and Triton give ? I have been stuck on the same setup and workflow for like 4 months :D

Cheers!

[Rogala]

SageAttention speeds up processing, and Triton is the library that allows it to do that.

[gonzalu]

Any downside? Is quality diminished in any way (even if insignificant) or is it still full fidelity just faster? I believe the TEa Cache thing is also a speed boost correct?

[Rogala]

You can use Sage Attention and TeaCache either separately or together, in any case there is a loss in quality sometimes insignificant sometimes very noticeable, acceleration is not magic but discarding some parts that are not really relevant or not important for the final result. But still you should just compare all these combinations to see if it suits you.

It's like mp3 for music, it has maximum compression, which is enough for a person to perceive the sound series, but it loses some parts that he can't hear anyway.

[Heer-Boaz]

pip install https://github.com/woct0rdho/SageAttention/releases/download/v2.1.1-windows/sageattention-2.1.1+cu128torch2.7.0-cp312-cp312-win_amd64.whl

Not part of the conversation and yet want to denote, for anyone facing the same problem, that running the command
pip install -v -U git+https://github.com/facebookresearch/xformers.git@main#egg=xformers --no-build-isolation
resolved the xformers error CUDA error (/__w/xformers/xformers/third_party/flash-attention/hopper/flash_fwd_launch_template.h:175): no kernel image is available for execution on the device.

Environment: Win11, WSL2, Ubuntu, RTX 5070 Ti, Git-cloned ComfyUI.

[euss3]

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
controlnet-aux 0.0.9 requires opencv-python-headless, which is not installed.
controlnet-aux 0.0.9 requires scikit-image, which is not installed.
controlnet-aux 0.0.9 requires timm<=0.6.7, which is not installed.
fvcore 0.1.5.post20221221 requires yacs>=0.1.6, which is not installed.
hfutils 0.11.0 requires click>=7, which is not installed.
pandas 2.2.3 requires python-dateutil>=2.8.2, which is not installed.
pycollada 0.9 requires python-dateutil>=2.2, which is not installed.

[euss3]

did you manged to run this workflow? drag and drop the video in comfui web: https://civitai.com/images/58389191

[Zzzeen2552]

After installed fresh comfyui using this standalone ComfyUI package with a cuda 12.8 torch build (https://github.com/comfyanonymous/ComfyUI/releases/download/latest/ComfyUI_cu128_50XX.7z)
Finally comfyui can run on mynew 5060Ti.

But, i have a new problem. The all my workflow collection is broken when i open it.

Right clicking : Fix nodes (recreate) didnt fix it. Need to replace it manually one by one.

Any idea how to solve this? Please help.

[Turkinolith]

Those are custom nodes, you'll need to check the repos for whoever built those. Looks like https://github.com/yorkane/ComfyUI-KYNode/tree/main

[KDB51]

May i ask for the status in regards to the desktop version, also running a 5090.
thank you in advance

[ssuukk]

For my ComfyUI I ditched the default python/pip/poetry (I guess?) installation and dependency mechanism and migrated to uv. It just works, either on Windows or on Linux. If anyone is interested, I can post my uv pyproject.toml somewhere. Everything works on 5090, couldn't get sageattention to work, though. Plus migration to other device/operating system/whatever is completely painless.

[hwprinz]

don't know, what the issue is on 50xx series meanwhile
just install the actual default portable package
pytorch2.7.0+cu128
done, all you need
then install Xformers from here: https://github.com/facebookresearch/xformers#installing-xformers
but use for portable:

# [linux & win] cuda 12.8 version
python -m pip install -U xformers --index-url https://download.pytorch.org/whl/cu128

then goto Tritron as it is needed for sage-attention
and download the lib package here https://github.com/woct0rdho/triton-windows?tab=readme-ov-file#8-special-notes-for-comfyui-with-embeded-python
the pytorch2.7.0 lib package put into python_embeded folder and 'extract Here'
then, as directly described as two line above

Note again that if you're using the embeded Python, then instead of directly run pip, you need:
C:\path\to\python_embeded\python.exe -m pip install -U "triton-windows<3.4"

it will pull and install latest tritron-windows 3.3.0post19
last: (for that actual COmfy Install) is sage-attention itself
read here:
just do

C:\path\to\python_embeded\python.exe -m pip install https://github.com/woct0rdho/SageAttention/releases/download/v2.1.1-windows/sageattention-2.1.1+cu128torch2.7.0-cp312-cp312-win_amd64.whl

(the full list of releases to be found: https://github.com/woct0rdho/SageAttention/releases)
as per this discussion some above #6643 (reply in thread)

lastly:
Add to the run_nvidia_gpu.bat file
--use-sage-attention --fast

You will end with a start like this

Checkpoint files will always be loaded safely.
Total VRAM 23027 MB, total RAM 32639 MB
pytorch version: 2.7.0+cu128
xformers version: 0.0.30+9a2cd3ef.d20250321
Enabled fp16 accumulation.
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 Ti : cudaMallocAsync
Using sage attention
Python version: 3.12.9 (tags/v3.12.9:fdb8142, Feb 4 2025, 15:27:58) [MSC v.1942 64 bit (AMD64)]
ComfyUI version: 0.3.30
ComfyUI frontend version: 1.17.11

Note: since sage-attention is 'stronger' than xformers, it will be used over xformers.
but if you start without '--use-sage-attention'
the line will say
Using xformers attention

all DONE and working

[ethanfel]

that's windows though

[Zzzeen2552]

After installing standalone comfyui 0.3.34 everything work well using RTX 4060Ti 16gb. I can run Wan VACE workflow, FLOAT lipsync etc,

But after installing xformers, triton and sageattention,some custom nodes getting error. Especially for video custom nodes like WanVideoWrapper, FramepackWrapper, comfyui-SONIC and comfyUI-FLOAT.

This is my ERROR log.

Checkpoint files will always be loaded safely.
Total VRAM 16311 MB, total RAM 65457 MB
pytorch version: 2.7.0+cu128
xformers version: 0.0.30
Enabled fp16 accumulation.
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 5060 Ti : cudaMallocAsync
Using sage attention
Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
ComfyUI version: 0.3.34
ComfyUI frontend version: 1.18.10
[Prompt Server] web root: C:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfyui_frontend_package\static
Traceback (most recent call last):
File "C:\ComfyUI_windows_portable\ComfyUI\nodes.py", line 2131, in load_custom_node
module_spec.loader.exec_module(module)
File "", line 999, in exec_module
File "", line 488, in _call_with_frames_removed
File "C:\ComfyUI_windows_portable\ComfyUI\comfy_api_nodes\canary.py", line 5, in
raise Exception("INSTALL NEW VERSION OF PYAV TO USE API NODES.")
Exception: INSTALL NEW VERSION OF PYAV TO USE API NODES.

Cannot import C:\ComfyUI_windows_portable\ComfyUI\comfy_api_nodes\canary.py module for custom nodes: INSTALL NEW VERSION OF PYAV TO USE API NODES.
C:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\windows_utils.py:315: UserWarning: Failed to find Python libs.
warnings.warn("Failed to find Python libs.")
C:/Users/ADMINI1/AppData/Local/Temp/tmpe5ol9g9u/cuda_utils.c:14: error: include file 'Python.h' not found
Failed to compile. cc_cmd: ['C:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\runtime\tcc\tcc.exe', 'C:\Users\ADMINI1\AppData\Local\Temp\tmpe5ol9g9u\cuda_utils.c', '-O3', '-shared', '-Wno-psabi', '-o', 'C:\Users\ADMINI1\AppData\Local\Temp\tmpe5ol9g9u\cuda_utils.cp312-win_amd64.pyd', '-fPIC', '-lcuda', '-lpython3', '-LC:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\backends\nvidia\lib', '-LC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\lib\x64', '-IC:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\backends\nvidia\include', '-IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\include', '-IC:\Users\ADMINI1\AppData\Local\Temp\tmpe5ol9g9u', '-IC:\ComfyUI_windows_portable\python_embeded\Include']
C:/Users/ADMINI1/AppData/Local/Temp/tmped17c4re/cuda_utils.c:14: error: include file 'Python.h' not found
Failed to compile. cc_cmd: ['C:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\runtime\tcc\tcc.exe', 'C:\Users\ADMINI1\AppData\Local\Temp\tmped17c4re\cuda_utils.c', '-O3', '-shared', '-Wno-psabi', '-o', 'C:\Users\ADMINI1\AppData\Local\Temp\tmped17c4re\cuda_utils.cp312-win_amd64.pyd', '-fPIC', '-lcuda', '-lpython3', '-LC:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\backends\nvidia\lib', '-LC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\lib\x64', '-IC:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\triton\backends\nvidia\include', '-IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\include', '-IC:\Users\ADMINI1\AppData\Local\Temp\tmped17c4re', '-IC:\ComfyUI_windows_portable\python_embeded\Include']

Anyone knows what happened?

Environment: Win11, RTX 5060 Ti, Git-cloned ComfyUI.

[Rogala]

A lot of things are broken, but you need to go to the python_embeded folder, open the console (cmd) and apply the updates to the packages:

.\python.exe -m pip install --upgrade pip
.\python.exe -m pip install -r .\ComfyUI\requirements.txt
.\python.exe -m pip install --upgrade torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu128

This may solve the problems, but the node packages may also need to be updated, and it is advisable to clear the cache in the browser on page 127.0.0.1:8188

[hwprinz]

Danger here....
they if you actually do the install based on what @Rogala wrote here ↑
right now you get installed torch2.7.1+cu128 and all accordingly....
which is not really a problem, but native Xformers 0.0.30 needs torch 2.7.0
the version i had in my log above 'xformers version: 0.0.30+9a2cd3ef.d20250321' ( find them here https://huggingface.co/Panchovix/xformers-windows-blackwell2.0-nightly/tree/main)
checks out torch>=2.6.0
so, already having 2.7.0 works, upgrading to torch2.7.1 should still work as well

i haven't upgraded to 2.7.1 yet, because i'm fine with 2.7.0

[yu-jingxiang]

Message ID: ***@***.***> Wow, you can use Sage Attention. I have tried installing it many times but it has failed. I am using CUDA128+Python 2.7, as well as the two wheels I sent you, xformers version: 0.0.30+9a2cd3ef.d20250321。 There is currently the latest version of Python 2.7.1, but I haven't tried it yet because I need too many things to support their operation. Too tired. 从网易163邮箱发来的超大附件推荐客户端极速下载 python.rar (322.67M, 2025年6月21日 14:40 到期) 下载

[hwprinz]

just follow my 'installguide' above
ten minutes job, but ensure to stay on torch 2.7.0, because there is not yet any prebuild whl for torch2.7.1
#6643 (reply in thread)

[yu-jingxiang]

Hello, I don't know why I'm updating all Triton 3.3.1 xformer 0.0.30 and installing Sageattention, but it still reports the same error [2025-06-06 15:18:06.851] [info] Set attention implementation to nunchaku-fp16 Loading configuration from E:\models\diffusion_models\svdq-fp4-flux.1-dev\comfy_config.json model_type FLUX Requested to load CLIPVisionModelProjection loaded completely 6937.49091796875 787.7150573730469 True CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16 clip missing: ['text_projection.weight'] Waiting for your reply Message ID: ***@***.***>

[hwprinz]

at least i don't see an erroe here ↑ in this
btw, 'Triton 3.3.1' and tritron-windows (for sage-attention) are different pair of shoes... just sayin'
and on all three accelerators: Xformers, Sage-attention, Flash-attention
they are used on different purposes
and myself i have never explicitly seen a message that explicitly sage or flash attention is getting used
the only real mesage i have seen so far is when xfomers is used instead of pytorch on vae

but you can test if sage-attention is present and active... there are loads of possibilities
or just try a SD1.5 workflow. it will rain messages when sage-attention is active....
they don't like each other

[yu-jingxiang]

Thank you, friend. My previous problem has been solved, My gguf and nunkaku can both run. But there is an old problem, [2025-06-10 06:31:08.647] [info] Initializing QuantizedFluxModel on device 0 [2025-06-10 06:31:08.681] [info] Loading weights from E:\models\diffusion_models\svdq-fp4-flux.1-dev\transformer_blocks.safetensors [2025-06-10 06:31:08.685] [warning] Failed to load safetensors using method MIO: CUDA error: operation not supported (at C:\Users\muyang\Desktop\nunchaku-dev\src\Serialization.cpp:130) [2025-06-10 06:31:18.948] [info] Done. Injecting quantized module [2025-06-10 06:31:19.296] [info] Set attention implementation to nunchaku-fp16 Loading configuration from E:\models\diffusion_models\svdq-fp4-flux.1-dev\comfy_config.json model_type FLUX Requested to load CLIPVisionModelProjection loaded completely 6937.49091796875 787.7150573730469 True CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16 clip missing: ['text_projection.weight'] It's always like this CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16 clip missing: ['text_projection.weight']， "clip missing: ['text_projection.weight']" What exactly does this mean It can also be generated in the end, why does it keep prompting this! Have you ever seen a similar problem? At 2025-05-15 22:08:22, "H.W.Prinz" ***@***.***> wrote: @a1chera did you get it fixed? i had the same issue, and it is not solvable from my point of view atm (already had some discussion with Dr. LT.Data on it) the Nighty portable packages come with python 3.13.2 that is the only problem you can install a bunch of nodes and stuff not taking care on it, but some, even essential as like Manager don't like 3.13..... at all. but if you're going for an actual download default released package, it will have python 3.12.10 pytorch2.7.0+cu12.8 all good with runs flawlessly have fun — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>